From LPMeissner@msn.com Thu Dec 5 16:41:49 1996 Received: from upsmot01.msn.com (upsmot01.msn.com [204.95.110.78]) by dkuug.dk (8.6.12/8.6.12) with ESMTP id QAA06330 for ; Thu, 5 Dec 1996 16:41:46 +0100 Received: from upmajb04.msn.com ([204.95.110.81]) by upsmot01.msn.com (8.6.8.1/Configuration 4) with SMTP id HAA01677; Thu, 5 Dec 1996 07:40:16 -0800 Date: Thu, 5 Dec 96 15:41:04 UT From: "Loren Meissner" Message-Id: To: "R. Baker Kearfott" Cc: sc22wg5@dkuug.dk Subject: RE: (SC22WG5.1212) Proposed Industrial Strength Standard Fortran Subset The concept of "sequence association" is described in 12.4.1.4 of Fortran 95. For example, the actual argument can be an array element when the dummy argument is an array. Also, an actual argument can be an array of strings when the dummy argument is one string. The storage sequence for the dummy argument string is taken from the actual argument string array in "standard array element order" as though the element strings of the actual array were concatenated together. ========================================== If all interfaces are required to be explicit, this kind of argument mismatching is eliminated anyway. My proposal would not prohibit passing arguments by reference, but your example would have to be written differently. A "preferred" F90/F95 version (which is acceptable according to my proposal) would be written more like this: call Simulated_Dynamic( A(: N), A(N+1: 2*N), A(2*N+1: 3*N) ) . . . subroutine Simulated_Dynamic( A, B, C ) real, dimension (:) :: A, B, C ! Assumed shape arrays integer :: N . . . N = size( A ) . . . In this example, the use of array sections as actual arguments (instead of array elements) is the important difference. My proposal would not prohibit passing N as an actual argument and declaring A, B, and C with size N, as in your example - the assumed shapes are just a style preference. - - Loren Meissner ---------- From: R. Baker Kearfott Sent: Wednesday, December 04, 1996 6:30 AM To: Loren Meissner Subject: Re: (SC22WG5.1212) Proposed Industrial Strength Standard Fortran Subset Loren, I, too, like the proposal, except for one question, as I note below. Baker At 12:27 AM 12/4/96 UT, Loren Meissner wrote: >WHITE PAPER >by Loren Meissner, 3 December 1996 > >Objectives: > - Get rid of deprecated features with a "fresh start" instead of > incrementally > - Enforce the safety features that are optional in F90 and F95 > - Do not put unnecessary roadblocks in the way of conversion from full > standard Fortran to the Subset language > >1. Remove storage association. This includes COMMON, EQUIVALENCE, and >argument (array/string) sequence association. This is obviously the >"biggie". No subset that retains storage association is worth the effort. > I have no problem with COMMON and EQUIVALENCE in this subset. But I don't fully understand removal of array/string sequence association. Does that mean that one could not pass arguments by reference? Does that mean that one can't have CALL SIMULATED_DYNAMIC(N,A(1),A(N+1),A(2*N+1)) SUBROUTINE SIMULATED_DYNAMIC(N,A,B,C) DIMENSION A(N), B(N), C(N) ? Best regards, Baker --------------------------------------------------------------- R. Baker Kearfott, rbk@usl.edu (318) 482-5346 (fax) (318) 482-5270 (work) (318) 981-9744 (home) URL: http://interval.usl.edu/kearfott.html Department of Mathematics, University of Southwestern Louisiana USL Box 4-1010, Lafayette, LA 70504-1010, USA ---------------------------------------------------------------